一种基于知识图谱的技术功效图自动构建方法  

A Method of Constructing Patent Technology-Effect Map Automatically Based on Knowledge Map

作　　者：张兆锋[1,2] 张均胜[2] 姚长青[2] 

机构地区：[1]南京大学信息管理学院,江苏南京210093 [2]中国科学技术信息研究所,北京100038

出　　处：《情报理论与实践》2018年第3期149-155,共7页Information Studies:Theory & Application

基　　金：国家自然科学基金项目"科学基金项目产出专利对产业技术发展的影响研究"(项目编号:L1624039);国家自然科学基金项目"基于文献的科研事件语义链接及其应用研究"(项目编号:71503240)的成果

摘　　要：文章研究了一种基于知识图谱进行技术功效图自动构建的方法。[方法/过程]具体包括专利文本中技术主题、应用领域、功效等知识抽取方法,知识图谱构建技术,技术功效图自动构建模型、数据流程等,并对新能源汽车领域专利进行实证研究。[结果/结论]实现了基于知识图谱的技术功效图自动构建,减少了人工参与,提高了技术功效图构建的速度、交互性和灵活性。[局限]该方法较大依赖于词系统中术语与术语间关系的丰富程度。[Purpose/significance] This paper proposes an automatic approach to construct Technology-Effect Map from patent text based on knowledge map. [Method/process]The approach includes knowledge extraction methods such as the themes,applications,and effect of technology in patent texts,knowledge graph construction technology,Technology-Effect Map automatic construction model and workflow. The patents from new energy automobile domain are used for empirical study. [Result/conclusion]The experiment shows that the approach could improve the efficiency,interactivity and flexibility of Technology-Effect Map by reducing manual work of human being. [Limitations] The method relies heavily on the richness of the relation among terminologies in the word system.

关 键 词：知识图谱 技术功效图 专利分析 自动构建 

分 类 号：G306[文化科学]